Immigrant background peer effects in Italian schools.

This article provides an empirical assessment of the effect of the concentration of students of immigrant origin on student learning, in Italian primary and lower secondary schools. I draw on the data of a national standardized learning assessment administered in 2010 to the entire student population at selected grades. The main threat to identification is given by the endogeneity of school characteristics, due to the fact that families choose their children's schools. To circumvent this problem I exploit the within-school random variability observed in the share of immigrant students across classes. I estimate peer effects allowing for heterogeneous effects between native and immigrant background children, and among natives, between children of different socio-economic background. The main finding is that the proportion of children of immigrant origin has a weak negative effect on child learning outcomes. This negative effect is somewhat larger for children of immigrant and low socioeconomic background, while it is negligible or even positive for high social origin native children.